When an engineering PhD engages in the pig farming industry
China is the world's largest producer and consumer of live pigs, with news related to pigs always attracting special attention, from selling pork at stalls set up by graduates of prestigious schools to the 8888 yuan per night "Pig View Room".
At the "Hello: Made in China Zhi" exhibition currently held at the China Industrial Design Museum located in Songnan, Baoshan, a booth related to pigs has also sparked discussions among visitors. The exhibition traces the historical process of the transformation from "Made in China" to "Made in China". The "Transformation and Rise of China's Intelligent Manufacturing" section vividly showcases China's transformation of traditional industries with new design concepts, digital technology, and intelligent technology through over a hundred advanced equipment models in China's sea, land, air, and aerospace fields, as well as digital products from scientific and technological innovation enterprises. The digitization of pig farming is one of the highlights.
Through exhibitions, people can see that pigs also have "ID cards", and farmers can observe the growth of pigs in real time through live cameras... How can the artificial intelligence Internet of Things empower pig farming and solve pain points in various links of the industrial chain? What is the market prospect of "intelligent pig farming" and what changes can it bring to the industry? The reporter talked to Lan Song, a Ph.D. in engineering and chairman of Rui Animal Science and Technology, about pig farming.
Cracking the pain points of human labor counting pigs
There are jokes circulating on social media that fund managers in suits and leather shoes personally go to pig farms to count pigs in order to take stock of company assets; The urban beauties who audit at the four major accounting firms need to "bathe and change clothes" before going to the pig farm to count pigs, and then soak in the "ocean of pigs" for several consecutive days.
It may sound unbelievable, but industry insiders say it's all true. Not only can outsiders not figure out how many pigs there are in the pig farm, but farmers themselves may not be able to give an accurate number. Since it is inevitable to count the quantity, is there a more intelligent and accurate means to replace manpower?
Rui Animal Technology has provided a solution. Through intelligent traceability products and artificial intelligence IoT platforms, the number and status of pigs can be easily identified, and the full lifecycle management of pigs from weaning to slaughter can also be achieved. At present, the digitalization and standardization level of the domestic breeding industry is still at a relatively low stage, and the digital pig breeding market is still a blue ocean. As a result, Rui Animal Technology has developed rapidly and has accumulated customers in more than 20 provinces and cities nationwide since 2017, serving millions of pigs.
Lan Song introduced that the focus is on the full lifecycle management of live pigs, aiming to solve the actual pain points of the pig farming industry. For example, the health and growth issues of pigs. The breeding industry is most afraid of diseases, but traditional pig farms mostly judge based on experience, and subjective errors are inevitable. However, Rui Animal Technology can monitor the physiological index of live pigs in real time through intelligent sensing systems. After wearing temperature measuring devices, the temperature data of live pigs will be automatically collected and uploaded to the cloud. After data processing by AI algorithm models, it will output judgments. Once sick pigs are detected, they can be immediately detected to prevent the spread of the epidemic.
Whether pigs grow well or not used to be judged only by their eyes and scales. With the development of a breeding status analysis system based on computer vision and Internet of Things technology, cameras collect real-time image data from the pigsty, which identifies the individual and behavioral characteristics of the pigs and intelligently calculates the average weight of the pigsty. In other words, farmers have a clear understanding of how much each pig eats per day, how well it is digested and absorbed, and how its weight curve changes. The quality of pig farming is closely related to the environment of the pig farm. By relying on temperature and humidity sensors, the system can also achieve environmental monitoring, and once there is a numerical abnormality, it will immediately issue a warning to the user through a mobile phone.
Building industry data infrastructure
Why did an engineering PhD devote himself to the pig farming industry? Upon hearing this question, Lan Song explained with a smile, "Pigs and pig farms are service recipients. Essentially, we are a technology company that provides digital solutions."
In 2015, Lan Song was preparing for entrepreneurship while pursuing a PhD in Integrated Circuit Design at Nanyang Technological University in Singapore. Due to his family's involvement in animal husbandry and his own technical background, Lan Song aimed to conduct feasibility studies on the digitalization direction of the pig farming industry. After the establishment of Rui Animal Technology, Lan Song decided to return to China to form a team. At that time, we inspected several first tier cities and ultimately chose to establish a research and development center in Shanghai, where talent, technology, capital and other factors were most densely concentrated. Nowadays, Rui Nu Technology has more than 30 office research and development teams in Shanghai, all of whom graduated from well-known domestic and foreign universities in fields such as data analysis and computer science.
To do product development, it is necessary to have a deep understanding of application scenarios. In the first few years of entrepreneurship, Rui Animal Technology invested a lot of energy in data accumulation work. For this reason, the company has established a pig farm in Nantong, Jiangsu. System engineers and algorithm engineers have come forward to weigh the pigs, repeatedly conduct data training, and improve the accuracy of camera calculation of pig weight. "Only by accumulating a large amount of data, labels and training can we gradually polish the model and improve the accuracy. It is also through this hard work stage that we now have the amount of data accumulation that many large Internet companies do not have." Lan Song said that in view of the large group customers raising many varieties of pigs, enterprises have made relevant training models for different varieties of pigs in the system. At present, the self-developed "Tianpeng System" by the enterprise has an inventory accuracy of nearly 100% and an estimated weight accuracy of 98%, which is at the leading level in the industry.
After inspecting large-scale breeding farms across the country, Rui Animal Science and Technology found that the generally poor digital foundation of the breeding industry poses challenges to the application of technology. Lan Song said, "In some remote areas, there are problems with the network environment and even power supply of pig farms, so we need to find ways to solve the problem of stable data transmission and no loss." Therefore, the R&D team designed functions for harsh factors in complex scenarios, such as the pig farm needing to be regularly flushed with a high-pressure water gun. The team developed edge servers with IP67 waterproof rating, which ensure that data can be cached to local storage first without a network, and also have power storage function, which can work for 8-10 hours in case of power outage.
At present, Rui Animal Technology is striving towards the goal of making traditional industries innovative and efficient by building industry data infrastructure, expanding service content and paths, connecting more upstream and downstream links in the industrial chain. "We will also continue to root in Shanghai, work with more partners, and let more people understand and value the value of digitalization in terms of funding, breeding efficiency, management efficiency, and other aspects," said Lan Song.
